Documents
Resources
Learning Center
Upload
Plans & pricing Sign in
Sign Out

Ashok - PowerPoint

VIEWS: 24 PAGES: 20

									                                  INDIA – CMS, 5-6 April 2007




    Search for the Standard Model Higgs Boson
 Decaying to Muons via Weak Gauge Boson Fusion


                                   Ashok Kumar and Suman Beri
                                  Panjab University – Chandigarh

                                        Ritva Kinnunen
                                   Helsinki Institute of Physics



                                                                   Ashok Kumar
INDIA – CMS, PU, 5-6 April 2007
                                                                   PU & GNDU
                          Cross-sections and Branching Fractions

Though the H → m + m - branching fraction is tiny and the backgrounds from Z→m + m -
and ttbar production are large, but it is an interesting channel because it provides
information on the Hmm coupling




                                             s[pb]
  Two search methods possible:
     - Inclusive search profiting from the larger cross section, fast simulation
      study in CMS IN 2005/013
    - Search in the weak gauge boson fusion, exploits an efficient background
       reduction with forward jet tagging
                                                                      Ashok Kumar
 INDIA – CMS, PU, 5-6 April 2007
                                                                      PU & GNDU
                                  Event Generation and Reconstruction

                                           generator   reconstructed   s x BR     s x BR x
                                                        events used             preselection
           Signal, mH = 120 GeV             PYTHIA         8000        1.1 fb      1.1 fb

        Z+jet, 10 < p T < 100 GeV1)         PYTHIA       288413        5.1 nb     5.1 nb

     tt, preselection cuts on muons 2)      PYTHIA       268244        840 pb     25.1 pb
     qqZ, preselection cuts on jets 3)     MadGraph       10000                    76 fb


1)     Z+jet generation with qq’ -> gZ and gq -> qZ,
       preselection: 10 < pT < 100 GeV, Z-> mm, ee, tt

2) preselection cuts: W-> mn, pTm1 > 20 GeV, pTm2 > 10 GeV ,|hm|<2.5, Mmm > 10 GeV

3) preselection cuts: pTq > 20 GeV, |hq|<5, |hq1 –hq2| > 4.2, hq1 * hq2 < 0,
                    Mq1,q2 > 900 GeV, Mmm > 70 GeV



 Standard ORCA packages for reconstruction:
 - Muons: GlobalMuonReconstructor
- Hadronic jets: TransientJetAlgorithm with Jet reconstruction cone = 0.5, Corrections: MCJet with
SplittedEcalPlusHcalTowerInput, Ettower > 0.5 GeV, Etower > 0.8 GeV and with ET(raw jet) > 20 GeV
- MET from EcalPlusHcalTowers with Type1 jet corrections
                                                                                Ashok Kumar
     INDIA – CMS, PU, 5-6 April 2007
                                                                                PU & GNDU
                                   Event Selection
                           Higgs boson mass reconstruction
Select events with exactly two isolated muons with pTm > 20 GeV, passing the L1 and
HLT di-muon or single muon triggers
Muon isolation: no tracks with pT>1 GeV in a cone of DR = 0.4 around the muon

New values due to the potential bb+jets background, pTm >15 GeV and DR =
0.3 were used in the first version

                    MC muons                          Reconstructed muons




 Results from a gaussian fit: s = 1.42 GeV (1.18%), <m mm > = 120.1 GeV
                                                                     Ashok Kumar
 INDIA – CMS, PU, 5-6 April 2007
                                                                     PU & GNDU
                                  Central variables for event selection
   1. No neutrinos in the signal event: an upper bound on ETmiss can be used to
      suppress the ttbar background
   2. Expect more boosted di-muon system in the signal than in the Z+jet background:
                mm
      cut in pT




                                                                                        mm
  Rejection of tt ~4 for ETmiss < 50 GeV               Rejection of Z+jets ~12 for pT >60 GeV
                                                                          Ashok Kumar
INDIA – CMS, PU, 5-6 April 2007
                                                                          PU & GNDU
               Number of signal and background events and cut efficiencies
                       after central event cuts and mass window
                                    Signal           qqZ             Z+jets                     ttbar

   Events for 60 fb -1                66            5160            3.05 x 10 8              1.25 x 10 6
            L1                    61.8 (93.5%)   5142 (99.7%)   1.18 x 10 8 (38.7%)      1.25 x 10 6(100%)
           HLT                    57.5 (93.0%)   4930 (95.9%)   4.84 x 10 7 (41.0%)     1.00 x 10 6(80.2%)
            PV                    56.6 (98.5%)   4863 (98.6%)   4.74 x 10 7 (97.8%)     9.97 x 10 5(99.4%)
    2 m,pTm>20 GeV                41.9 (74.4%)   3668 (75.4%)   2.09 x 10 7 (44.1%)     3.79 x 10 5(38.0%)
    muon isolation                28.8 (68.9%)   2546 (69.4%)   1.44 x 10 7 (68.9%)     1.60 x 10 5(42.2%)
    MET < 50 GeV                  25.7 (89.2%)   2179 (85.6%)   1.37 x 10 7 (95.1%)     4.26 x 104(26.6%)
     pTmm > 60 GeV                16.6 (64.4%)   1396 (64.1%)   1.10 x 10 6 (8.0%)      2.27 x 10 4(53.3%)
118 < mmm <122 GeV                11.1 (67.0%)   2.06 (0.15%)     2968 (0.27%)             383.6 (1.69%)



No signal visibility possible even with large integrated
luminosity, assume that Higgs boson mass is known from
other processes to select the mass window

                                                                                      Ashok Kumar
INDIA – CMS, PU, 5-6 April 2007
                                                                                      PU & GNDU
           Forward jet tagging in the mass window1)

Selection of forward jets: two hardest jets anywhere in rapidity with h j1*h j2 < 0


Select lowest possible value,
ET > 20 GeV,
due to PYTHIA generation of Z+jet
background
No optimization of this variable


Signal vs qqZ differences due to
generator level cuts on qqZ
Differences may arise also due to the possibility
of the Z originating from a final state quark


1) Due to the small mass window, statistics low for Z+jets:
   - need to open the mass window and
   - assume decoupling of pTmm and tagging cuts
                                                                     Ashok Kumar
INDIA – CMS, PU, 5-6 April 2007
                                                                     PU & GNDU
                           Rapidity gap between the tagging jets

       Expect large rapidity gap for the q jets in qq -> qqH




      |hq1 –hq2| > 4.2 at the generator level
      for qqZ
                                                        |h j1 - h j2| > 4.5


                                                                         Ashok Kumar
INDIA – CMS, PU, 5-6 April 2007
                                                                         PU & GNDU
                                  Di-jet invariant mass

            With all (previous) event selection cuts, including |hj1 - hj2| > 4.5




                                                  Optimal cut value: Mtagging jets > 1 TeV

                                                                           Ashok Kumar
INDIA – CMS, PU, 5-6 April 2007
                                                                           PU & GNDU
                                  Central jet veto

Selection of central jets:

- All corrected jets between the two tagging jets within
hmin + 0.5 < hjet < hmax – 0.5 and with E T > 20 GeV



 Statistics too low for Z+jet
 to obtain the efficiency curve

 ETjet = 20 GeV taken as the
 threshold




                                                           Ashok Kumar
INDIA – CMS, PU, 5-6 April 2007
                                                           PU & GNDU
           Number of signal and background events and cut efficiencies
               for forward jet tagging and central jet veto cuts

                                    Signal          qqZ            Z+jets                ttbar

   Events for 60 fb -1                66            5160         3.05 x 10 8          1.25 x 10 6
Central event cuts,               11.1 (67.0%)   2.06 (0.15%)   2968 (0.27%)        383.6 (1.69%)
118 < mmm <122 GeV
     2 tagging jets               8.5 (76.7%)    1.88 (91.3%)   1033 (34.8%)        162.8 (42.4%)
   ETj1,ETj2>20 GeV               7.7 (90.0%)    1.81 (96.5%)   373 (36.1%)         149.9 (92.1%)
      |hj1-hj2|> 4.5              4.0 (51.9%)    1.52 (83.8%)   52.6 (14.1%)          7.7 (5.1%)
       mjj > 1 TeV                2.4 (59.0%)    1.36 (89.2%)    4.1 (7.7%)          1.8 (24.0%)
    central jet veto              1.7 (69.0%)    0.95 (69.9%)   2.3 (56.3%)          0.5 (27.8%)




                                                                               Ashok Kumar
INDIA – CMS, PU, 5-6 April 2007
                                                                               PU & GNDU
                    Comparison of PYTHIA and ALPGEN generation
                             for the Z+jets background

Data samples:
- existing ALPGEN production samples for exclusive 2-5 jet final states,
generated with Z-> ee, mm,tt, mll > 40 GeV

- generation of PYTHIA events with the data cards used for the ORCA
production samples but with mll > 40 GeV


Comparison method:
Jet reconstruction within |h | < 5 with the PYTHIA jet reconstruction
package in CMKIN
Main event selection cuts:
  - two leptons (electrons or muons) with p T > 15 GeV/c
 - cut in the di-lepton pT, pTll > 60 GeV/c
 - two tagging jets with the method used for the ORCA analysis
  - central jet veto
  - No MET cut, no mass window cut

                                                          Ashok Kumar
  INDIA – CMS, PU, 5-6 April 2007
                                                          PU & GNDU
                                                   ALPGEN generation



                                     2j       3j          4j         5j       2j-5j          PYTHIA

s x BR (pb)                         230      66           17        4.0        317            4500

                                    105      30          7.8        1.8        145             831
2l, pT > 15GeV/c
pTll > 60 GeV/c                     35       14          4.1        1.1        54              60

2 jets, hj1*hj2<0                   10       5.0         1.6        0.5        17              19

    |hj1-hj2|> 4.5                  0.16    0.14       3.4x10-2   1.5x10-2    0.35            0.12

     mjj > 1 TeV              4.7x10-2     5.1x10-2    1.4x10-2   1.2x10-2    0.12           6.7x10-2

ET(veto)< 20 GeV              8.4x10-3     5.3x10-3    3.6x10-4   2.6x10-4   1.4x10-2        2.2x10-2




                                                                               Ashok Kumar
  INDIA – CMS, PU, 5-6 April 2007
                                                                               PU & GNDU
               Conclusion from ALPGEN/PYTHIA comparison


                                     ALPGEN           PYTHIA
efficiency of jet tagging cuts      2.2 x 10-3       1.1 x 10-3

 efficiency of central jet veto       12%               33%


Differencies due to harder jets in ALPGEN


- Higher efficiency from forward jet tagging cuts from ALPGEN,
- higher efficiency for central jet veto from PYTHIA

+ normalization in PYTHIA (pT > 10 GeV/c)

-> level of Z+jets background in ORCA simulation not underestimated




                                                       Ashok Kumar
INDIA – CMS, PU, 5-6 April 2007
                                                       PU & GNDU
            Estimate of WW+jets, W->mn background

 Exploit the ALPGEN samples:
  s(WW+2jets) = 1.22 pb
 - PYTHIA jet reconstruction
 - Main event relection (as for Z+2jet events)
 - Result: s(selected) = 6.5x10-5 pb, 0.8% of Z+2jet cross section



                       Estimate of bb+jets background

  Estimate based on the study of bb background on bbH/A->em+X
  PTDR2, page 346, Table 11.1:
  - Leptonic triggers, lepton pT cuts (pTm>19 GeV, pTe>29 GeV) and
  isolation (pT > 1 GeV, DR = 0.4): 37.4 pb, rejection factor ~106
  - Assume that for the remaining selection the efficiency is similar to
  that of Z+jets
  -Result: ~ 4x10-3 events in the mass window



                                                           Ashok Kumar
INDIA – CMS, PU, 5-6 April 2007
                                                           PU & GNDU
                     Signal superimposed on the total background

                              With all selection cuts, for 60 fb -1




                                                                      Ashok Kumar
INDIA – CMS, PU, 5-6 April 2007
                                                                      PU & GNDU
                   JES uncertainty in the background determination

 Signal visibility not obtained with 60 fb -1 , background could be estimated
 around the already known Higgs mass (from H->gg)
 Main uncertainty expected from JES uncertainty

  Uncertainty on the backgrounds from JES uncertainty,
  10% MET uncertainty included

                                                        tt most sensitive to the
                                                        increase of JES uncertainty
                                                        due to large jet multiplicity


                                                         30% uncertainty in this
                                                         estimate due to MC statistics




                                                                       Ashok Kumar
INDIA – CMS, PU, 5-6 April 2007
                                                                       PU & GNDU
                        Estimate of statistical significance


        Total background in the mass bin 3.74 events

        Systematic uncertainties:
         JES uncertainty dominates for this channel
           tt 20%
           Z+jets 16%
           qqZ 8%
         Assume 5% uncertainty for all production cross sections

        Total background with systematics 3.74 ± 0.73 events

        Significance (Poisson stat., ScPf program):
         0.75 with systematic uncertainty
         0.80 without systematic uncertainty



                                                               Ashok Kumar
INDIA – CMS, PU, 5-6 April 2007
                                                               PU & GNDU
                                  Conclusions
 Full simulation study on qq → qqH, H → mm with forward jet tagging and
 central jet veto yields S/B ~ 0.5 in a mass window assuming m H to be known
 from H → gg for 60 fb-1

 Statistical significance = 0.75 and upper bound for sxBR = 12.7 fb at 95% CL,
 including systematic uncertainty

 Main backgrounds from Z+jets, ttbar and qqZ were simulated


 For better optimization of jet tagging and veto cuts, generation of the Z+jets
 background with a NLO generator needed

 For high luminosity, Lt > 100 fb-1, inclusive H->mm channel more profitable:
 background can be extracted with high precision around the known mass value
 (0.5% from fast simulation for 60 fb-1)


                                                                      Ashok Kumar
INDIA – CMS, PU, 5-6 April 2007
                                                                      PU & GNDU
  The analysis note was presented in the Physics meeting on
  December 8, 2006 by Suman Beri

  Minutes of the Physics Meeting of 08/12/06:
  AN2006/105 Search for the Standard Model Higgs Boson Decaying into Muons

  - Comments from room: approach of the analysis is not the right one.
  presently this is approached like a discovery analysis, while there is no
  chance in this analysis. Instead one should take into account the
  already measured mass of the Higgs, select on a mass window
  in the di-muon system and then optimize the cuts against background.
  In any case it will remain a difficult channel for the LHC.

  - Z+jets ALPGEN files are now available so these files can be used to
  make the background analysis of that channel, or at least cross check
  explicitely against PYTHIA

  - Concerns from the referees are to be taken to be taken into account

   STATUS: Not approved. Sent back to the Higgs group to re-discuss the
   analysis strategy

                                                               Ashok Kumar
INDIA – CMS, PU, 5-6 April 2007
                                                               PU & GNDU

								
To top